Antelog — Safety layer for AI agents
Runtime security for AI agents
Stop AI agents before they break things.
Antelog checks every action AI agents take on your developers’ computers, and stops the ones that would destroy or leak your data, before they run.
Installation commands
macOS and Linux$ curl -fsSL https://github.com/YashSarwaiya/antelog-releases/releases/latest/download/install.sh | sh
Windows (PowerShell)$ irm https://github.com/YashSarwaiya/antelog-releases/releases/latest/download/install.ps1 | iex
Incident log
This already happened.
- GIGAZINERead story (opens in a new tab)
Data loss: CursorAI agent deletes PocketOS’s database and every backup in one call. With Antelog: Stopped. - Tom’s HardwareRead story (opens in a new tab)
Data loss: Claude Code runs terraform destroy on a live site. With Antelog: Stopped. - SnykRead story (opens in a new tab)
Secrets: Malware turns AI coding tools into password hunters. With Antelog: Stopped. - FortuneRead story (opens in a new tab)
Data loss: Replit agent wipes a live database during a code freeze. With Antelog: Stopped.
How it works
Every action is checked first.
- Log it
Saves the command before anything runs. - Check it
Safety rules look for anything dangerous.
- Safe: it runs
$ npm test- Dangerous: stopped
$ DROP DATABASE prod;
- Dangerous: stopped
Fast. Each check takes under 0.5 seconds on average.
Coverage
Disasters stop here. Your work doesn’t.
- Antelog checks every action before it runs.
- The AI can’t turn Antelog off.
- Antelog stays quiet during normal work.
Leak
- Prompt injection: Antelog blocks sending data where a web page said to, uploading files, emailing data, etc.
- Secrets: AI tries to show or send your keys. Blocked on the spot (e.g., printing the .env file, AWS credentials).
- Personal info: AI tries to send customer data outside. Blocked before it leaves (e.g., customer emails, copied customer records).
Destroy
- Data and backups: AI tries to wipe live data or its backups. Antelog stops it first (e.g., deleting the production database).
- Cloud: AI tries to tear down live servers. Blocked before anything breaks (e.g., deleting a cloud database).
- Code and files: AI tries to erase your work. It never goes through (e.g., deleting GitHub repository).
Install
Start protection with one command.
Terminal
- macOS and Linux
curl -fsSL https://github.com/YashSarwaiya/antelog-releases/releases/latest/download/install.sh | sh - Windows (PowerShell)
irm https://github.com/YashSarwaiya/antelog-releases/releases/latest/download/install.ps1 | iex
Pricing
Free to start.
FreeWas $20 now $0
For one person. Runs on your computer.
- Blocks data leaks and data destruction.
- Online dashboard.
- 50,000 AI checks a month per machine.
- 14-day online history.
Team
Custom
For companies whose AI agents can reach real systems and data.
- Everything in Free, on every computer.
- Company rules and company places on every computer.
- Audit log and CSV export.
FAQ
How do I stop Claude Code or Cursor from running dangerous commands?
Install Antelog to check every command your agent wants to run.
Can it stop an AI agent from deleting my production database?
Yes, it stops destructive commands like DROP TABLE or DELETE.
What doesn’t it protect against?
Commands you run yourself and agents Antelog isn’t connected to.
Can my agent still delete test data?
Yes, dropping test databases is never stopped.
Does it slow my agent down?
Rule checks run on your computer and add milliseconds.
Get started
Give your agent a seatbelt.
- Your code agent: Claude Code, Cursor.
- A shield: rules + AI checker.
- A check: only checked actions run.
- antelog: every action checked before it runs.